Microsimulation (a.k.a. microanalytic simulation) is a modelling technique that operates at the level of individual units such as persons, households, vehicles or firms. Within the model each unit is represented by a record containing a unique identifier and a set of associated attributes – e.g. a list of persons with known age, sex, marital and employment status; or a list of vehicles with known origins, destinations and operational characteristics. A set of rules (transition probabilities) are then applied to these units leading to simulated changes in state and behaviour. These rules may be deterministic (probability = 1), such as changes in tax liability resulting from changes in tax regulations, or stochastic (probability <=1), such as chance of dying, marrying, giving birth or moving within a given time period. In either case the result is an estimate of the outcomes of applying these rules, possibly over many time steps, including both total overall aggregate change and, crucially, the distributional nature of any change.

Uses of microsimulation

Given the emphasis on changes in distribution, microsimulation models are often used to investigate the impacts on social equity of fiscal and demographic changes (and their interactions). Modelling of the distribution of traffic flows over a street network is another increasingly important use of the approach.

Other individual-level modelling approaches

Microsimulation (MSM) is closely allied to two other individual-level modelling approaches: Cellular Automata (CAs) and Agent Based Models (ABMs). In their originally conceived forms, these three approaches may be regarded as representing the three corners of a continuum of individual level modelling approaches.

In a pure CA all entities are spatially located within a grid of cells, and all entities have only one attribute (alive or dead), with behaviours deterministically dependent upon the state of neighbouring cells. In a pure ABM the emphasis is on the interaction between individuals, with the main attribute of each individual being their operating characteristics (behavioural rules), which evolve stochastically over time in response to the success or failure of interactions with other individuals. In a pure MSM transition probabilities lack evolutionary and spatial dimensions. As microsimulation models add more behavioural and spatial interaction between individual units, as CAs add a growing range of individual attributes and start to incorporate aspatial behaviours, and as ABMs add both space and fiscal/demographic characteristics to their agents, the three approaches move towards a common ground.